Creation of explanatory sentences for a series of pictures by analyzing their relation

In this paper, a system that generates explanatory sentences for a series of pictures is proposed. In much the same way as humans do, this system evaluates object emotion, the spatial relationship between objects, object behavior and the connective relationships between pictures. It does this by using objective information such as facial expressions, and the position and size of objects. Our system applies soft computing techniques such as neural network models, fuzzy reasoning, and case-based reasoning. Its effectiveness is also demonstrated by simulation.